Why is clock speed not good measure of system performance ?

Question

why is clock speed not good measure of system performance ??
and can you tell me what is better and explain

Answer 1

A CPU is just one part of the system. For most common tasks a Pentium 4 2GHZ cpu can manage with 10% of its processing power and the chart will show just 10%.

Most activities don't need much of CPU so other things matter a lot. For example someone using a P3 CPU but with a 52X speed cd writer will be much faster than someone using a p4 PC with a 4X speed writer.

In case you are surfing on a dial P4 then you will much slower than someone on an ADSL link but a P3 cpu.

Taking printing. Someone with a p3 CPU but a 16ppm laser can print 16 pages per minute. Someone on a P4 will print only 2 pages on a normal inkjet.

Then there are things like ram, hard disk speed. Most Desktops use faster parts (but use more energy) and therefore can beat a laptop with a faster CPU.

Answer 2

These days so many other factors influence program performance that simply using clock speed is not proper and it never was for comparisons between different processors - e.g. the Apple IIe used a 1 MHz 6502 and was widely considered faster that a 4 MHz 8086 in the IBM PC running Visicalc because of the way the two processors used the system clock - basically the 8086 stepped across instructions while the 6502 used the leading and trailing edges.
Today, the data bus runs at a different speed than the processor and subsystems preload memory into faster cache memory to have it ready for the fast processor to use. Two systems identical except for the amount of cache can show great differences in processing, especially if the task is doing calculations or searches on lots of data. And if the hard disk has differing amounts of cache for loading data off the drive disks, that may affect things a lot.
There when a system is running more than one job they may interfere with each other.
Because of the bloated nature of many tasks now popular on PC's including data searching, graphics, video, music, and games, the more cache memory available at all levels - hard drive, video, processor - and the more memory over all available, the "better" the system will be.

Answer 3

Have you ever heard the saying a chain is only as strong as its weakest link? That principle applies here. Think of your PC as a car. If I wanted to build a really fast car that handles well and is extremely comfortable I would get the best of every component. But still there are limits. The clock speed of a processor would be like the RPM's of a combustion engine. A 4-cylinder engine might rev to 10,000 RPMs but only make 200-HP. A 12-cylinder engine may rev to 8,000 RPM's and make 800-HP. All the components of a PC determine its speed. Same as with a car. If you drop a V-12 in where a 4-cylinder was you can brag about your engine, but unless you upgrade the transmission, it probably won't make much difference. Similarly if you upgrade a processor without upgrading the processor end of the bus (RAM speed, mobo capabilities) it probably won't do as much as you think. Any more, video has such a huge impact on on everything we do, you're most cost effective upgrade is the video card. There are applications out there for benchmarking your system and seeing where a system needs help...and giving a general number relating to the overall speed of your system. Check them out at http://madonion.com There are others out there that will actually compare your system against classic systems, like a Pentium 75 with 64M of ram and tell you how many times faster yours is. Building a fast PC is like building a fast car, you have to upgrade everything to do it right.

Answer 4

For the same reason a cars stated speed is not a good measure of how long it will take to get from point A to point B. There are too many other potential sources of delay to use that stat alone. Every major component in a computer system is a potential bottleneck. A fast CPU will end up waiting on a slow video card, hard drive, system bus, software application etc.

Don't be surprised I know this stuff - Pimps are high -tech these days.

Answer 5

Clock speed is important, but it's not all you have to consider. The number of processors (single, dual, or quad) plays a big part in performance, but even that can be misleading if the processing you want to do can't be divided among multiply processors (single threaded vs. multi-threaded). The speed of the front side bus (FSB) is also important, but if you don't have much cache or RAM, your processing power will be reduced. Some processors are designed to optimize floating point processing (SPECfp), while others are designed to optimize integer processing (SPECint). The bottom line is there are a lot of factors that influence a system's performance, and clock speed is only one of them.

Answer 6

Clock speed does not tell you how much work a processor can do. It depends on how many clock cycles does it take to perform an instruction? If it has parallel logic then it might do several instructions at the same time. If it is multi core it might be able to multiply the calculating power without needing to increase the clock speed.

Answer 7

The escapement in a pendulum governed clock releases spring pressure to the pendulum drive. Each time the escapement of a clock clicks one tooth it imparts a small amount of spring pressure to overcome the friction of the pendulum bearings and the pendulums wind resistance. All pendulum clocks run faster freshly wound and slow down as the spring pressure diminishes toward the end of a winding. This is usually only a few seconds a day and can be corrected for over the time that the winding lasts by adjusting the length of the pendulum. Increasing the weight too much can cause extra friction so that the spring tension released by the escapement may not be enough to sustain the pendulums swing. If you keep the weight close to the original weight of the pendulum you should still be able to adjust the overall clock accuracy with the length of the pendulum. Just provide a means of moving the sprocket up and down for adjustment and keep the wind resistance to a minimum by reducing the frontal area of the sprocket and its mounting.

Answer 8

System performance is best measured, in my opinion, by the number of processors (i.e., single, dual, quad) plus memory (I have 4 GB on my setup, which has made an enormous difference). Regarding clock speed, "overclocking" is notorious for rendering systems unstable.

Answer 9

because it isn't the clock or clock-chopper that slows down the rest of the circuit. I think data bus speed might be a better measure of system performance.

Answer 10

no one factor is a good measure of system performance, many things should be considered, such as operating system, amount of RAM, clock speed and number of processors, system cache, onboard video memory etc etc..

another thing to consider, what the systems primary function will be.